A list of professions most vulnerable to artificial intelligence sparked widespread controversy after Andrei Karpathy, one of the founders of OpenAI, published it and later deleted it due to a misunderstanding of its content.
Karpathy, who contributed to the development of the ChatGPT model, prepared this list using artificial intelligence techniques to analyze data on occupations in the “Occupational Outlook Guide” issued by the US Bureau of Labor Statistics (BLS), a guide that includes about 143 million jobs in various sectors of the economy.
The list relied on giving each job a score from 0 to 10 to assess the extent to which it was “exposed to the influence of artificial intelligence,” with higher scores indicating a greater likelihood of replacing the job or integrating artificial intelligence tools into its tasks.
The results revealed that high-paying jobs were the most vulnerable to the impact, while low-income jobs, especially those with salaries of less than $35,000 annually, were the least affected. Software developers, data scientists and financial analysts topped the list of jobs most at risk, while professions such as construction workers, barbers and nursing assistants were among the least at risk.
But Karpathy quickly deleted the list from his website, explaining that many people misinterpreted it as a direct prediction of the future of jobs. He wrote in a post on the X platform that the goal of publishing the data is to give researchers and developers the opportunity to visually explore it and perform various analyzes on it.
He stressed that the “degree of exposure” determined by the Large Linguistic Model (LLM) depends primarily on the extent of the digital nature of the job, and does not necessarily reflect what will actually happen in the labor market, noting that the fate of professions is linked to other factors such as the nature of demand and economic conditions. (Russia Today)
